Measurement of Nitrophenols in Rain andAir by Two-Dimensional LiquidChromatography-Chemically Active Liquid CoreWaveguide Spectrometry

Lucksagoon Ganranoo,†,⊥ Santosh K. Mishra,† Abul K. Azad,† Ado Shigihara,†

Purnendu K. Dasgupta,*,† Zachary S. Breitbach,† Daniel W. Armstrong,† Kate Grudpan,‡ andBernhard Rappenglueck§

Department of Chemistry and Biochemistry, The University of Texas at Arlington, Arlington, Texas 76019-0065,Department of Chemistry, Chiang Mai University, Chiang Mai 50200, Thailand, and Department of Earth andAtmospheric Sciences, University of Houston, Houston, Texas

We report a novel system to analyze atmospheric nitrophe-nols (NPs). Rain or air sample extracts (1 mL) are precon-centrated on a narrow bore (2 mm) aliphatic anion ex-changer. In the absence of strong retention of NPs exhibitedby aromatic ion exchangers, retained NPs are eluted as aplug by injection of 100 µL of 0.1 M Na2SO4 on to a short(2 × 50 mm) reverse phase C-18 column packed with2.2 µm particles. The salt plug passes through the C-18column unretained while the NPs are separated by anammonium acetate buffered methanol-water eluent,compatible with mass spectrometry (MS). The elutedNPs are measured with a long path Teflon AF-basedliquid core waveguide (0.15 × 1420 mm) illuminatedby a 403 nm light emitting diode and detected by amonolithic photodiode-operational amplifier. The wave-guide is rendered chemically active by suspending it overconcentrated ammonia that permeates into the lumen.The NPs ionize to the yellow anion form (λmax ∼ 400 nm).The separation of 4-nitrophenol, 2,4-dinitrophenol,2-methyl-4-nitrophenol, 3-methyl-4-nitrophenol, and 2-ni-trophenol (these are the dominant NPs, typically in thatorder, in both rain and air of Houston and Arlington, TX,confirmed by tandem MS) takes just over 5 min withrespective S/N ) 3 limits of detection (LODs) of 60, 12,30, 67, and 23 pg/mL compared to MS/MS LODs of20, 49, 11, 20, and 210 pg/mL. Illustrative air and raindata are presented.

Environmental occurrence of nitrophenols (NPs) was firstreported 35 years ago;1,2 their toxicity to plants and humans hassince led to continued attention.3-6 Many NPs (especially 3-meth-yl-4-nitrophenol (3-Me-4-NP)) are potent vasodilators;7 2-NP can

reduce the oxygen carrying capacity of blood;8 4-NP and 2,4-dinitrophenol (2,4-DNP) are cytotoxic, mutagenic, and likelycarcinogenic.9 4-NP, routinely found in diesel exhaust aerosols,is both estrogenic and antiandrogenic.10 Many NPs cause head-ache, shortness of breath, nausea, fever, and peripheral numb-ness.1 Some NPs affect plant cell metabolism even at a 10 µMconcentration by uncoupling oxidative phosphorylation;11 somedecrease transpiration12 and inhibit uptake of nutrients13 andgrowth.14 Phytotoxicity of NPs may contribute to forest decline;11

compared to healthy leaves, higher NP levels are found ondamaged ones.15 NPs are also toxic to algae and fish.16

Atmospheric NPs have both primary and secondary sources.8

Primary sources include industrial raw materials or intermediatesinvolved in the manufacturing of plastics, pharmaceuticals, dis-infectants, dyes, explosives, herbicides, and insecticides; thesources stem from the use or disposal of these products,combustion of coal and wood, and notably, internal combustionengine exhaust.3,17 Secondary formation of NPs occurs atmo-spherically via sequential reactions of aromatics with •OH and

* Corresponding author. E-mail: Dasgupta@uta.edu.† The University of Texas at Arlington.‡ Chiang Mai University.§ University of Houston.⊥ Permanent address: Department of Chemistry, Chiang Mai University,

Chiang Mai 50200, Thailand.(1) Nojima, K.; Fukaya, K.; Fukui, S.; Kanno, S. Chemosphere 1975, 2, 77–82.(2) Nojima, K.; Fukaya, K.; Fukui, S.; Kanno, S.; Nishiyama, S.; Wade, Y.

Chemosphere 1976, 5, 25–30.

(3) Nojima, K.; Kawaguchi, A.; Ohya, T.; Kanno, S.; Hirobe, M. Chem. Pharm.Bull. 1983, 31, 1047–1051.

(4) Grosjean, D. Environ. Sci. Technol. 1985, 19, 968–974.(5) Leuenberger, C.; Llgocki, M. P.; Pankow, J. F. Environ. Sci. Technol. 1985,

19, 1053–1058.(6) Hinkel, M.; Reischl, A.; Schramm, K.-W.; Trautner, F.; Reissinger, M.;

Hutzinger, O. Chemosphere 1989, 18, 2433–2439.(7) Mori, Y.; Kamata, K.; Toda, N.; Hayashi, H.; Seki, K.; Taneda, S.; Yoshino,

S.; Sakushima, A.; Sakata, M.; Suzuki, A. K. Biol. Pharm. Bull. 2003, 26,394–395.

(8) Harrison, M. A. J.; Barra, S.; Borghesi, D.; Vione, D.; Arsene, C.; Olariu,R. I. Atmos. Environ. 2005, 39, 231–248.

(9) Huang, Q.; Wang, L.; Han, S. Chemosphere 1995, 30, 915–923.(10) Li, C. M.; Taneda, S.; Suzuki, A. K.; Furuta, C.; Watanabe, G.; Taya., K.

Toxicol. Appl. Pharmacol. 2006, 217, 1–6.(11) Rippen, G.; Zietz, E.; Frank, R.; Knacker, T.; Klopffer, W. Environ. Technol.

Lett. 1987, 8, 475–482.(12) Barber, D. A.; Koontz, H. V. Plant Physiol. 1963, 38, 60–65.(13) Votrubova-Vaoousova, O. Biol. Plant. 1977, 19, 166–172.(14) Bonner, J. Am. J. Bot. 1949, 36, 323–332.(15) Natangelo, M.; Mangiapan, S.; Bagnati, R.; Benfenati, E.; Fanelli, R.

Chemosphere 1999, 38, 1495–1503.(16) Howe, G. E.; Marking, L. L.; Bills, T. D.; Rach, J. J., Jr. Environ. Toxicol.

Chem. 1994, 13, 51–66.(17) Tremp, J.; Mattrel, P.; Fingler, S.; Giger., W. Water Air Soil Pollut. 1993,

68, 113–123.

Anal. Chem. 2010, 82, 5838–5843

10.1021/ac101015y 2010 American Chemical Society5838 Analytical Chemistry, Vol. 82, No. 13, July 1, 2010Published on Web 06/02/2010

NO2 in the daytime and NO3• and NO2 in the nighttime.18-20

In the aqueous phase, NPs can be formed from phenols bydissolved HONO photolysis.21 A comparison of atmosphericbenzene, toluene, xylenes, and ethylbenzene in air and thecorresponding NPs in rain suggest more rapid conversion ofalkylbenzenes to NPs than benzene,22 as would be expected fromtheir •OH reactivities.

Atmospheric NPs occur in aqueous, gaseous, and particulateforms. Concentrations of ng-µg/m3 have been reported;15,23-27 theyare distributed globally, and this includes remote regions.15,24,27

Among atmospheric NPs, 4-NP, 2-NP, and 2,4-DNP are reportedlythe major contributors. Our interest in atmospheric NPs lies in thatNPs may act as reservoirs for HONO. Much as they may form fromHONO by photolysis,21 they also release HONO by photolysis overa broad range of wavelengths.28 A lot is unclear about why morningozone peaks are observed in some major metropolitan areas,29 andNPs, via HONO, may provide the missing •OH source whereasHONO concentration by itself is inadequate. In this context, extantmeasurements of NP virtually all come from Europe; limitedmeasurements made in the U.S. are now 25 years old.5,30

MEASUREMENT STRATEGIES

In the following, the general literature on the measurement ofNPs is briefly surveyed. The majority of the extant approachesrely on extraction followed by chromatography-MS. More recentstrategies describe adsorption on XAD-7 sorbent, elution withsome organic solvent and analysis by liquid chromatography-atmospheric pressure chemical ionization mass spectrometry (LC-APCI-MS),31 extraction with polyacrylate solid phase microex-traction fibers, conversion to the t-butyldimethylsilyl derivativefollowed by GC/MS,32 liquid-liquid extraction, and LC-APCI-MS/MS;22 and solid phase extraction26,33 or supported liquid mem-brane microextraction34 followed by LC-UV. Using a dynamic

extractant film of n-butyl chloride in octanol, Miro et al.35 wasable to more sensitively measure several NPs simultaneously bymultiwavelength spectrometry and multiple linear regression thanin an earlier similar effort.36 Luttke et al.24 separated cloudwaterwith an impactor and collected the NPs in the effluent gas by amist chamber. They processed both the cloudwater and the gasphase extract by acidification, NaCl addition, sorption by a C18

sorbent, elution by ethyl acetate, and analysis by isotopedilution GC/MS. Hoffmann et al.37 concentrated NPs incloudwater and rainwater into methyl t-butyl ether (MTBE)by a novel extraction apparatus and carried out LC-ion trap MS/MSn. Leuenberger et al.5 collected aerosols on glass fiber filtersand (semi-) volatiles on a downstream polyurethane foam plugand Tenax-GC for gaseous constituents alone, followed bysolvent extraction/thermal desorption GC/MS. Hinkel et al.6

measured NPs on plant surfaces by acidic dichloromethaneextraction, diazomethane derivatization, and GC/MS.

Other methods, from enzyme-linked immunoassay (ELISA38)to differential pulse voltammetry on various electrodes,39,40 haveall been used. Amperometry has been used in the batch41,42 orthe flow injection mode43 and following capillary LC44 or electro-phoretic45 separation. 4-NP is often used as a tag in ELISA wherean alkaline phosphatase cleaves a p-nitrophenyl phosphate linkage;several of the above approaches target the liberated 4-NP formeasurement. One approach measures the loss of fluorescenceas 4-NP begins to absorb radiation around 400 nm that otherwiseexcites a fluorescent coumarin dye.46

The approaches above are not configured for automated fielduse. Most are not sensitive or robust enough for the purpose.Some amperometric methods report excellent sensitivity butrequire too frequent electrode maintenance in practice. Chroma-tography coupled to mass spectrometry, especially MS/MStechniques, have excellent sensitivity and selectivity but are hardlyreadily fieldable. MS-based approaches provide unequivocalconfirmation of identity but accurate quantitation requires expen-sive isotopically labeled standards. Considerable capital, operating,and maintenance costs further deter wide application.

Prior experience in designing and fielding instruments tomeasure soluble gases/particles with liquid/ion chromatographyapproaches47,48 predisposed us to a similar approach, especiallyas NPs, like inorganic salts, are water-soluble. However, unlike

(18) Atkinson, R.; Aschmann, S. M.; Arey., J. Environ. Sci. Technol. 1992, 26,1397–1403.

(19) Bolzacchini, E.; Bruschi, M.; Hjorth, J.; Meinardi, S.; Orlandi, M.; Rindone,B.; Rosenbohm, E. Environ. Sci. Technol. 2001, 35, 1791–1797.

(20) Olariu, R. I.; Klotz, B.; Barnes, I.; Becker, K. H.; Mocanu, R. Atmos. Environ.2002, 36, 3685–3697.

(21) Vione, D.; Maurino, V.; Minero, C.; Pelizzetti, E. Chemosphere 2001, 45,893–902.

(22) Kohler, M.; Heeb, N. V. Anal. Chem. 2003, 75, 3115–3121.(23) Leuenberger, D.; Czuczwa, J.; Tremp, J.; Giger, W. Chemosphere 1988,

17, 511–515.(24) Luttke, J.; Scheer, V.; Levsen, K.; Wunsch, G.; Cape, J. N.; Hargreaves,

K. J.; Storeton-West, R. L.; Acker, K.; Wieprecht, W.; Jones, B. Atmos.Environ. 1997, 31, 2637–2648.

(25) Wennrich, L.; Efer, J.; Englewood, W. Chromatographia 1995, 41, 361–366.

(26) Belloli, R.; Barletta, B.; Bolzacchini, E.; Meinardi, S.; Orlandi, M.; Rindone,B. J. Chromatogr., A 1999, 846, 277–281.

(27) Vanni, A.; Pellegrino, V.; Gamberini, R.; Calabria, A. Int. J. Environ. Anal.Chem. 2001, 79, 349–365.

(28) Bejan, I.; Abd El Aal, Y.; Barnes, I.; Benter, T.; Bohn, B.; Wiesen, P.;Kleffmann, J. Phys. Chem. Chem. Phys. 2006, 8, 2028–2035.

(29) Rappengluck, B.; Dasgupta, P. K.; Leuchner, M.; Li, Q.-Y.; Luke, W. Atmos.Chem. Phys. 2010, 10, 2413–2424.

(30) Kawamura, K.; Kaplan, I. R. Atmos. Environ. 1986, 20, 115–124.(31) Belloli, R.; Bolzacchini, E.; Clerici, L.; Rindone, B.; Sesana, G.; Librando,

V. Environ. Eng. Sci. 2006, 23, 405–415.(32) Jaber, F.; Schummer, C.; Chami, J. A.; Mirabel, P.; Millet, M. Anal. Bioanal.

Chem. 2007, 387, 2527–2535.(33) Schussler, W.; Nitschke, L. Chemosphere 2001, 42, 277–283.(34) Bishop, E. J.; Mitra, S. Anal. Chim. Acta 2007, 583, 10–14.

(35) Miro, M.; Cladera, A.; Estela, J. M.; Cerda, V. Anal. Chim. Acta 2001,438, 103–116.

(36) Cladera, A.; Miro, M.; Estela, J. M.; Cerda, V. Anal. Chim. Acta 2000,421, 155–166.

(37) Hofmann, D.; Hartmann, F.; Herrmann, H. Anal. Bioanal. Chem. 2008,391, 161–169.

(38) Oubina, A.; Ballesteros, B.; Galve, R.; Barcelo, D.; Marcoa, M.-P. Anal. Chim.Acta 1999, 387, 255–266.

(39) Ni, Y.; Wang, L.; Kokot, S. Anal. Chim. Acta 2001, 431, 101–113.(40) Rodrıguez, I. N.; Leyva, J. A. M.; Hidalgo, J. L.; de Cisneros, H. Anal. Chim.

Acta 1997, 344, 167–173.(41) Fanjul-Bolado, P.; Gonzalez-Garcia, M. B.; Costa-Garcia, A. Anal. Bioanal.

Chem. 2006, 385, 1202–1208.(42) Honeychurch, K. C.; Hart, J. P. Electroanalysis 2007, 19, 2176–2184.(43) Mendez, J. H.; Martinez, R. C.; Gonzalo, E. R.; Trancon, J. P. Anal. Chim.

Acta 1990, 228, 317–321.(44) Ruban, V. F. J. HRC CC 1993, 16, 663–665.(45) Fischer, J.; Barek, J.; Wang, J. Electroanalysis 2006, 18, 195–199.(46) Paliwal, S.; Wales, M.; Good, T.; Grimsley, J.; Wild, J.; Simonian, A. Anal.

Chim. Acta 2007, 596, 9–15.(47) Toda, K.; Dasgupta, P. K. Compr. Anal. Chem. 2008, 54, 639–683.(48) Ullah, S. M. R.; Takeuchi, M.; Dasgupta, P. K. Environ. Sci. Technol. 2006,

40, 962–968.

5839Analytical Chemistry, Vol. 82, No. 13, July 1, 2010

suppressed conductometry for ionic solutes, there are no selectivedetection schemes for NPs. We considered that like nitroanilines(basis of the Hammett acidity scale49), all NPs are also acid-baseindicators, except that they are weaker acids than nitroanilines.NPs turn yellow in base and absorb strongly around ∼400 nm;few other water-soluble atmospheric compounds absorb in thevisible region. A long path liquid core waveguide (LCW) cell mayprovide sufficient absorbance detection sensitivity;50 to providefurther selectivity, we sought to exploit the dual nature of NPs(ionic: can be preconcentrated on anion exchange media; hydro-phobic: can be separated on reverse phase media) and, thus,provide a 2-dimensional separation. Aqueous NPs can be precon-centrated on anion exchangers, eluted as a plug by salt to a reversephase column where separation can be conducted with a standardelectrospray ionization/mass spectrometry (ESI/MS)-compatibleeluent. Instead of ESI/MS, however, LCW-based absorbance canbe routinely measured after base introduction to ensure ionizationof NPs. To further identify NPs, it will be possible to performdetection with and without basification while the ESI/MS-compatible eluent will allow MS confirmation without methodchangeover. We describe the results of such an approach.

EXPERIMENTAL SECTIONReagents and Materials. These were obtained from standard

suppliers; please see Supporting Information.Rain Collection. Rain samples were collected using a sequen-

tial wet-only rain sampler (Rain-go-round, www.horiba.com) thatcollects rain in 5 mL increments through a 46.5 cm2 opening;this provides a ∼1.1 mm resolution between samples. Thesampler was set in an open area at the University of Texas atArlington, ∼9.2 m above ground level, away from any roofprojections. Samples were stored at -20 °C until analysis.

Air Sampling. Sampling cartridges consisted of 1 g 35-60mesh silica gel in three-eighths in. i.d. polypropylene tubes withglass wool end retainers; terminal barbed fittings for one-eighthin. i.d. tubes were capped after filling. Poly(tertrafluoro)ethylene(PTFE) tubes were used throughout.

A Labview-controlled 24-channel sampler was used for sam-pling through an inverted glass funnel inlet at 2 standard liters/

min (SLPM) sequentially through each cartridge for 1 h; the mainmanifold flow was 20 L/min to minimize line losses. The samplerwas located atop a 17-story building on the University of Houstoncampus (Houston, TX) in April to May of 2009. Every 24 h,cartridges were replaced and sampled ones capped and stored insealed bags at -20 °C until analysis.

Standard Gaseous Nitrophenol Calibration Source. We madepermeation tubes individually filled with solid 2-NP and 5-MeO-2-NP and gravimetrically calibrated them. Known concentrationstreams of adjustable humidity were generated by dilution; seedescription and Figures S1 and S2 in the Supporting Information.

Sample Preparation. Frozen rain samples were allowed toreach room temperature and filled in autosampler vials through0.45 µm pore size nylon syringe filters; some were spiked inaddition to check recovery. Frozen sampled cartridges were keptat room temperature for 2 h and then were eluted with 2 mL of50% v/v methanol. The eluate was diluted to 4 mL with Milli-Qwater. This ∼<25% methanolic solution was filtered and loadedinto autosampler vials. Standard stock (100 mg/L) NP solutionsprepared in 10 mM Na2CO3 were diluted serially with 25% v/vmethanol to prepare working standards. Milli-Q water and fieldblank cartridges were, respectively, used as controls for rainand air samples.

Two-Dimensional Preconcentration and Separation. Thegeneral arrangement shown in Figure 1 uses a chromatographicpump (GS-50, this and all columns from www.dionex.com) and aprogrammable syringe pump (V6, www.kloehn.com). The latterdelivered 1 mL of sample at 0.48 mL/min from the autosamplerto an anion exchanger preconcentrator (AG21, 2 × 50 mm)contained within the loop of 6-port injector V1 (shown in theloading mode). Meanwhile, the 100 µL loop of a second 6-portvalve V2 is filled with an eluent plug of 100 mM Na2SO4 in 4 mMNH4OAc/42% v/v MeOH. Both V1 and V2 are switchedsimultaneously. The chromatographic pump delivers the saltplug from V2 to the preconcentration column, and the sampleis eluted to the reverse phase separation column (Acclaim-120C-18, 2.2 µm, 2.1 × 50 mm) that separates the solutes by adifferent retention mechanism. The isocratic eluent programuses 4 mM NH4OAc/42% v/v MeOH. As real samples wereanalyzed, we incorporated a wash step with methanol prior tosample injection. A 10-port valve was used for V2, and using asecond loop, 250 µL of methanol was injected prior to any

(49) Laitinen, H. A.; Harris, W. E. Chemical Analysis, 2nd ed.; McGraw-Hill: NewYork, 1975; pp 90-92.

(50) Dallas, T.; Dasgupta, P. K. TrAC, Trends Anal. Chem. 2004, 23, 385–392.

Figure 1. General experimental setup.

5840 Analytical Chemistry, Vol. 82, No. 13, July 1, 2010

analysis. Later, since the pump was gradient capable, we useda binary eluent with A ) 10 mM aqueous NH4OAc and B )70% v/v MeOH/H2O; normally, an isocratic composition of 60%B was used; except at the beginning of each analytical cycle, atriangular pulse of B (t ) 0f1f2 min, %B ) 60f100f60) wasused to remove strongly retained components. The analyticalcycle, operated under control of Peaknet V6.0 software, was<10 min, and could be completed in as little as 7 min.

Chemically Active Liquid Core Waveguide (CA-LCW) Detec-tor. Figure 2 shows the general arrangement. A 0.15 mm i.d × 1420mm long Teflon AF capillary is coupled to a 403 nm light emittingdiode LED at one end; the other end of the AF tube was coupled tothe hemispherical ball lens of a photodiode. Details are given in theSupporting Information; liquid inlet/outlet connections were madethrough the LED and the photodiode coupling. All but 10 cm at eachend of the AF tube was put within a 50 mL capacity polypropylenejar containing ∼25 mL of concentrated ammonia at the bottom,diluted 1:1 with water. Ammonia transfers through the highlypermeable51,52 AF tube making the LCW chemically active andraising the pH of the internal solution.

Tandem Mass Spectrometry. We used a TSQ QuantumDiscovery Max triple-quadrupole mass spectrometer (www.thermo.com) in the negative electrospray ionization mode with a heatedelectrospray ionization probe maintained at 325 °C. The vaporizerand capillary temperatures were 270 and 350 °C, respectively, withan electrospray voltage of 4000 V. A programmed diversion valvewas used to direct the LC output to the MS only in desired timewindows. Fragmentation and monitoring details for specific analytesare given in the Supporting Information.

RESULTS AND DISCUSSIONSelectivity and Dimensionality in Preconcentration and

Separation. Too many compounds in air and rain absorb in theUV region. NPs, however, absorb strongly in the visible region,and few atmospheric compounds show such indicator behavior.Still, one would wish to preconcentrate NPs selectively beforeanalysis. As indicated, we investigated preconcentration by anionexchange, plug elution with a salt and separation by a differentretention mechanism. We initially chose separation columns that

provide both anion exchange and hydrophobic retention (PAX-100, PAX-500, WAX-1). With ESI/MS compatible eluents, mostNPs were retained too strongly. In contrast, reverse phase columnsprovided attractive separations; similar reports are in the literature(see introduction). We chose a short, narrow bore, microparticulate(2 × 50 mm, 2.2 µm) C-18 column that separated all the major NPsin air and rain at our location (vide infra) in ∼5 min.

Affecting plug elution of NPs preconcentrated by anionexchange proved difficult. Most anion exchangers have aromaticskeletons, and the π-π interaction with the NPs are too strongto accomplish plug elution with any reasonable concentration andvolume of a salt plug. We considered acrylate and polyvinylskeleton anion exchangers, but none were available as shortcolumns. Anion exchangers based on the amine-diepoxidechemistry53,54 were suggested to us. Up to 1 mL of the mixtureof test NPs spiked in a rain matrix or in 25% methanol (represent-ing air sampling cartridge eluate) was quantitatively captured bysuch a column (AG21, Dionex), and plug elution of these on to aC-18 column by 100 µL of 100 mM Na2SO4 was readily possible.The salt plug actually promoted the retention of the highly polarDNP on the C-18 column.

CA-LCW Detection Strategy. Often, chromatography/reac-tion is optimal at one pH, but detection is optimal at another.Earlier, we solved an analogous problem by introducing NH3

through a membrane prior to detection;55 other reagents havealso been introduced without dilution.56 With Teflon AF,analytes have been deliberately preconcentrated into the wallsof a long path cell57 or a volatile background removed;58 othermembranes have also been used for this latter purpose.59

Presently, an elevated pH improves detection by completingionization. However, raising the eluent pH would have affectedthe separation and likely adversely affected the column. Theunusual properties of Teflon AF offered a unique opportunityto directly permeate NH3 into the waveguide and render it“chemically active”.

The measured pH of the methanolic eluent is ∼6.7; theavailable pKa values stated below are in water,60 limiting theinterpretation. Nevertheless, the ionization of 2,4-DNP (pKa

3.94) will be complete, but that of 4-NP (pKa 7.08), 2-NP (pKa

7.23), 3-Me-4-NP (pKa 7.33), and 2-Me-4-NP (pKa not listed,should be similar to 3-Me-4-NP) will be incomplete. Forconditions where the absorption from the unionized NP isnegligible, it is readily shown (see Supporting Information) thatthe ratio (R) of the signals (Sfin and Sin) at the different pH levels(pHfin and pHin) is given by

R ) Sfin/Sin)(KIn+Hin)/(KIn+Hfin)... (1)

where KIn is the dissociation constant of the NP and Hin, Hfin

are the hydrogen ion concentrations corresponding respectively

(51) Pinnau, I.; Toy, L. G. J. Membr. Sci. 1996, 109, 125–133.(52) Dasgupta, P. K.; Genfa, Z.; Poruthoor, S. K.; Caldwell, S.; Dong, S.; Liu,

S.-Y. Anal. Chem. 1998, 70, 4661–4669.

(53) Pohl, C. A.; Saini, C. U.S. Patent 7,291,395. November 6, 2007.(54) Kuban, P.; Dasgupta, P. K.; Pohl, C. A. Anal. Chem. 2007, 79, 5462–5467.(55) Hwang, H.; Dasgupta, P. K. Anal. Chem. 1986, 58, 1521–1524.(56) Zhang, G.; Dasgupta, P. K.; Sigg, A. Anal. Chim. Acta 1992, 260, 57–64.(57) Larsson, H.; Dasgupta, P. K. Anal. Chim. Acta 2003, 485, 155–167.(58) Saari-Nordhaus, R.; Anderson, J. M., Jr. J. Chromatogr., A 2002, 956, 15–

22.(59) Ullah, SM. R.; Adams, R. L.; Srinivasan, K.; Dasgupta, P. K. Anal. Chem.

2004, 76, 7084–7093.(60) Schwarzenbach, R. P.; Stierli, R.; Folsom, B. R.; Zeyer, J. Environ. Sci.

Technol. 1998, 22, 83–92.

Figure 2. Chemically active liquid core waveguide absorbancedetector.

5841Analytical Chemistry, Vol. 82, No. 13, July 1, 2010

to pHin, pHfin. A chromatogram of several test NPs with andwithout ammonia introduction is shown in Figure 3. Theobserved ratios (in the order they elute) are 1, 1.8, 2.6, 3.2, and2.9. The order of the first four is consistent with the KIn values.One would predict that 2-Me-4-NP will have a pKa value in 42%methanol between those of 2-NP and 3-Me-4-NP.

Initially, we planned splitting the column effluent into two LCWdetectors, to measure one with NH3 introduction and measurethe other without (or with HCl introduction) and, thus, establishan added measure of selectivity/identification. Experience withactual samples indicated, however, that the preconcentration/elution strategy and the spectral window together provide suchselectivity that the observed signal after NH3 introduction wasentirely due to a specific NP. This was confirmed by MSidentification, rendering dual LCW detection superfluous.

The use of long path LCWs increase detection sensitivity, buttoo large an LCW volume will result in loss of efficiency. Thepresent case is complex, however, as discussed in the SupportingInformation (Figure S3, Table S1, and accompanying discussion),efficiency can actually increase with increased LCW length forsome analytes up to a point. Theoretical and practical consider-ations surrounding the optimum length/volume of an LCWdetector will be addressed elsewhere.

Calibration Behavior and Detection Limits. CA-LCW de-tected calibration chromatograms of the dominant NPs appear inFigure 4; the corresponding area-based calibration plot is shownin Figure S3 in the Supporting Information. On the basis of theS/N ) 3 criterion, the respective limits of detection (LODs) for2,4-DNP, 4-NP, 2-NP, 3-Me-4-NP, and 2-Me-4-NP were 12, 60, 23,67, and 30 pg/mL. Mass LODs are the same numerical values inpg and compare with MS/MS LODs of 49, 20, 210, 20, and 11 pg,respectively, under otherwise identical conditions. For comparison,Hoffmann et al. have reported MS/MS LODs of 50-100 pg forvarious NPs after MTBE extraction. It is remarkable that the CA-LCW containing ∼$25 of electronics is capable of providing LODs

within a factor of 4 of the MS/MS instrument and actually doesbetter in one case. The CA-LCW approach may also be applicableto humic-like substances (HULIS) in the atmosphere, of consider-able current interest; this class of compounds is also bestseparated by a two-dimensional strategy.61

Nitrophenols in Rain. Rainfall in Arlington, TX contained4-NP as the most abundant NP, in one case close to 5 µg/L. Theconcentration data (analysis of 22 samples selected from thoseover a 15 month period spanning 5/08-9/09) for 4-NP, 2,4-DNP,2-Me-4-NP, and 2-NP respectively, were (mean ± sd, range): 1.80± 1.30, 0.20-4.89; 0.39 ± 0.70, 0.04-3.31; 0.65 ± 0.59, 0-2.26; and0.35 ± 0.20, 0-0.69 µg/L. 3-Me-4-NP was detectable in somesamples but was below the limit of quantitation (∼3× LOD). Allsamples were analyzed at least in duplicate. The mean variancefor replicate analysis of 4-NP, 2,4-DNP, 2-Np and 2-Me-4-NP inindividual samples was 2.3, 6.8, 25.0, and 7.3%, increasing withdecreasing average analyte concentration.

In the five samples collected over 2/09-3/09 in Houston, 2-NPwas undetectable and 2-Me-4 NP (mean ± sd: 0.32 ± 0.24, range:0.15-0.49 µg/L) exceeded 2,4-DNP (0.17 ± 0.12, 0.05-0.29 µg/L). While 4-NP was still the dominant NP (1.03 ± 0.02, 0.82-1.24µg/L), 3-Me-4-NP was easily measurable in all samples (0.15 ±0.07, 0.09-0.22 µg/L). An illustrative MS/MS chromatogram ofa rain sample is shown in Figure S5 in the Supporting Information.Overall, measured NP concentrations in rain were in the samerange as those reported in Japan2 and Europe.23,33

3-Methyl-2-Nitrophenol vs 3-Methyl-4-Nitrophenol. It isdifficult to distinguish between these two isomers by reversephase separation; MS/MS is not able to easily resolve this either.However, they are readily separated by chiral phases thatrecognize molecular shapes.62,63 Such analysis showed that the

(61) Limbeck, A.; Handler, M.; Neuberger, B.; Klatzer, B.; Puxbaum, H. Anal.Chem. 2005, 77, 7288–7293.

(62) Armstrong, D. W.; Demond, W. J. Chromatogr. Sci. 1984, 22, 411–415.(63) Armstrong, D. W.; Demond, W.; Alak, A.; Hinze, W. L.; Riehl, T. E.; Bui,

K. H. Anal. Chem. 1985, 57, 234–237.

Figure 3. Chromatogram of a standard mixture of NPs with andwithout ammonia introduction through the LCW. Concentrations inorder of elution: 4, 8, 4, 8, and 4 ng/mL, respectively.

Figure 4. Calibration series for the dominant NPs.

5842 Analytical Chemistry, Vol. 82, No. 13, July 1, 2010

isomer is essentially exclusively 3-Me-4-NP. (See Figures S8-S10in the Supporting Information.)

Sequential Rain Samples. The relatively small samplerequirement allows sequential rainfall analysis with good (∼1 mm)resolution. The more soluble NPs are rather rapidly washed outfrom the atmosphere, and the cumulative deposition pattern as afunction of rainfall should fit a first order pattern. Figure 5 showssuch data for 10-21-2009 for three NPs. 3-Me-4-NP was detect-able in all samples but was too low to quantify. The best-fit firstorder removal rate constants for the washout pattern of 4-NP and2,4-DNP (the two most soluble NPs: Henry’s law constant, KH,7.7 × 104 and 1.2 × 104 M/atm, respectively17), are within 20%of each other. In contrast, 2-Me-4-NP exhibited an order ofmagnitude lower removal rate constant; KH is not available forthis compound but that for the 4-Me-2-NP isomer is 64 M/atm.It appears that sequential rain profiling can be a convenientmeasure of washout efficiencies of NPs and similar species.

Air Analysis. These results are discussed here only brieflybecause discrete sorbent sampling and off-line elution is not thebest way to measure gas phase NPs. Nevertheless, we participatedin a field campaign (http://sharp.hnet.uh.edu/doku.php/home)without sufficient time to develop a chromatographically interfacedcontinuous sampler. Belloli et al.26 previously found silica gel tobe superior to XAD-2 for trapping NPs. We compared XAD-4, XAD-16, and Na2CO3-impregnated silica gel as well. All XAD-sorbentseither had to be precleaned or bought precleaned. NPs trappedon Na2CO3-impregnated silica gel could not be eluted withwater/methanol/acetonitrile without acid addition; this neededneutralization prior to preconcentration. A 60 Å pore size silicagel appeared the best overall choice and was used. In a serialtwo-cartridge arrangement, the NP was entirely found in thefirst cartridge (sampling rate: 1 to 2 SLPM, 23-83% RH, and20-150 µg/m3 of 2-NP or 5-MeO-2-NP).

Initial experiments utilized pure methanol or acetonitrile forelution of a pure NP from the sorbent and were measured bydirect spectrometry after adding alkali. It was found, however,

that NPs cannot be concentrated on the AG-21 preconcentratorfrom pure methanol but 25% methanol was acceptable. We used50% methanol for elution and diluted 1:1 with water prior topreconcentration. Relevant data are presented in Table S2 in theSupporting Information.

Not surprisingly, the same NPs that we found to be dominantin rain were the dominant NPs in air. Figure 6 shows achromatogram of an air sample extract compared with an aqueousstandard, and the inset shows an illustration of the variation ofgaseous 4-NP over a 3-day period. Concentration minima werealways observed around the noon hour.

In summary, we have proposed a simple and novel techniquethat should be readily extendable to multiwavelength measure-ment and, thus, have broader potential applicability.

ACKNOWLEDGMENTThis work was supported by Houston Advanced Research

Center (Project H-114). We wish to acknowledge C. Pohl, K.Srinivasan. and S. M. R. Ullah for their help; the potentialapplicability of the amine-diepoxide-based preconcentrator wassuggested by C. Pohl. The assistance of L. Pedemonte and J.Golovko in changing cartridges and collecting rain samples isgratefully acknowledged. L.G. acknowledges the support from theThailand Research Fund through the Royal Golden Jubilee Ph.D.Program (Grant PHD/0224/2547).

SUPPORTING INFORMATION AVAILABLEAdditional information as noted in text. This material is

available free of charge via the Internet at http://pubs.acs.org.

Received for review April 18, 2010. Accepted May 20,2010.

AC101015Y

Figure 5. Triangles, circles, and squares represent cumulativedeposition of 4-NP (left ordinate), 2,4-DNP, and 2-Me-4-NP (rightordinate), respectively, as rainfall continues. Figure 6. Air sample extract chromatogram compared to an aqueous

standard containing 2,4-DNP (1 ng/mL), 4-NP (2 ng/mL), 2-NP (1 ng/mL), 3-Me-4-NP (2 ng/mL), and 2-methyl-4-nitrophenol (2-Me-4-NP, 1ng/mL). The inset shows air concentration data for 4-NP over a 3-dayperiod in Houston, TX; the minima always occurs around noon.

5843Analytical Chemistry, Vol. 82, No. 13, July 1, 2010